semiconeng,
I don't even know what to say to this one. Really Joe, you expect me to believe..... that YOU believe that Mr. EGO himself.... Jerry Sanders.... announced a "Worst Case Scenerio"???
I think this is the new and improved Jerry, who talks softly....
Re: 3.6 million K7s
I think this is if not worst case scenario, than at least extremely conservative estimate.
Let's just review the last quarter. AMD probably did not sell any Duron chips (while they expected to sell at least 200,000) and Tbird may have been the same story - at least the socket Tbird. Still, AMD achieved their plan - 1.8 million K7s - without any contribution from socket A CPUs that were "on the market" for 1 to 3 weeks of the quarter.
Now those unsold CPUs did not disappear (that's why I said that 3.6 million K7 would be disappointing), but AMD went half of the quarter with very poor chipset support for Socket A CPUs. But Jerry reiterated 3.6 million target. To me, it suggests that if the stars were lined up correctly (meaning given sufficient supplies of quality chipsets), AMD could sell a lot more than 3.6 million chips, and 3.6 is a conservative number which can be achieved even with all the problems present.
This is why 3.6 million chips is the wrong place to start the yield calculation.
Of course, we could sit here and speculate all day, it wouldn't prove anything. All I'm saying, is if AMD is at 30% capacity in Dresden, and Full Capacity in Texas, then their output numbers sound low..... Based On My Experience.
The situation at AMD is slightly different from that at Intel. AMD's challenge is to grow the market share in the mid to high performance segment, increase ASPs all while facing a shortage of chipsets.
It's a balancing act. Do you make more Tbirds, hoping for high ASPs, but pushing the CPUs to markets that have been dominated by Intel. Do you just go for volume with Duron in AMD's traditional value segment? Wait, you can't do that, not enough chipsets.
Should AMD produce flat out, knowing that there is a limit to how many CPUs they will be able to sell (limited by chipsets)? Or should they instead accelerate the transition in Dresden to Mustang core based chips, which will hit the market in Q4, when chipset problem will not be as severe?
In Austin, should they produce K6 based CPUs flat out (not making much money doing it, or should they convert more lines to .18u based Duron / Tbird - losing some production during the transition?
AMD is facing challenges, of which producing more chips is not the most difficult one to overcome.
Joe |
